Economics of Traffic Attraction by Transit Providers

Abstract

Internet transit is an economy where providers sell traffic-delivery services. Traffic attraction refers to BGP (Border Gateway Protocol) techniques enabling an AS (Autonomous System) to receive traffic that would otherwise flow elsewhere. Unlike prior studies that present security perspectives on traffic attraction or deal with economic considerations in game-theoretic settings, this paper focuses on the economics of customer-traffic attraction by transit providers and report extensive simulations in an Internet-scale model configured with realistic data on traffic, topology, and pricing. We consider traffic attraction by tier-1 tier-2, and tier-3 networks with 3 types of reactions by other net-works: filtering, customer-disconnections, and attempts of losing ASes to attract traffic to themselves. Our results demonstrate that transit providers can derive substantial financial benefits from attracting customer traffic, with tier-1 networks being in the strongest position to do so. The traffic attraction remains effective despite the countermeasures unless participation in them is broad.